Haematology and Blood Transfusion Vol. 26
Modern Trends in Human Leukemia IV
Edited by Neth, Gallo, Graf, Mannweiler, Winkler
© Springer-Verlag Berlin Heidelberg 1981
The Different Origin 01 Primary and Secondary Chromosome
Aberrations in Cancer
G. Levan and F. Mitelman
A. Abstract
B. Introduction
We have proposed a hypothetical model to
explain the role of chromosomal aberrations in
malignant development. In this model we
postulate two kinds of chromosomal changes:
(1) primary, active changes caused by direct
inter action between the oncogenic agent and
the hereditary material of the host cello These
changes are mainly somatic mutations, but
mayaIso be associated with directed structural
changes visible in the microscope; and (2)
secondary, passive changes arising randomly
by nondisjunction and structural rearrangements. They are followed by selection of cells
with changes that amplify the primary change
and thus appear as nonrandom chromosome
patterns.
This hypo thesis is discussed in the light of
1827 cases of human malignancy in which we
have recently surveyed and systematized chromosomal aberrations. Special support for the
idea of somatic mutations as the initiator of
malignant development comes from work of
Knudson and collaborators in human retinoblastoma. The Phl chromosome, predominant
during the chronic phase of chronic myeloid
leukemia (CML), is proposed as an instance of
a primary change, whereas the chromosome
changes during the blastic crisis of CML will
illustrate the secondary changes. The most
common of these secondary changes is actually
the doubling of the Phl and thus an amplification of the primary change. The increase
in number of copies of a specific chromosome reported by Green and collaborators
demonstrates that this kind of amplification
can result in direct response to the need
for a specific gene located in that chromosome.
Towards the end of the nineteenth century it
became known from the work of pathologists
that malignant cells were often liable to chromo so me aberrations. Boveri in 1914 hypothesized that these aberrations actually were the
cause of the malignant growth. At that time,
technical difficulties in the study of mammalian chromosomes effectively put a stop to any
thorough investigation of chromosome aberrations in tumors. Improvements in methodology in the early 1950s made detailed studies
feasible, but only after the introduction of
chromosome banding in the early 1970s could
chromosomal and subchromosomal changes in
tumor materials be analyzed with great precision. The results from both experimental and
human tumors were significant: c1ear nonrandom patterns of aberrations became discernible. There were, however, considerable difficulties in interpreting the findings. Firstly,
there were usually many different types of
changes even within a clinicaHy weH defined
group of tumors. Secondly, there was a variable degree of "background noise", i.e., chromosome changes that did not fall into the
nonrandom pattern. Thirdly, there were cases
of malignant tumors in which even refined
chromosome banding did not reveal any aberrations.
160
C. Chromosome Aberrations in Human
Malignancies
We have attempted to collect data as completely as possible from all human tumors that
have been studied carefully by chromosome
banding techniques. The bulk of the material
Myeloproliferative
disorders
Chronic myeloid leukemia (CML)
t(9 ;22) and other aberrations
Aberrant translocations
Acute myeloid leukemia (AML)
Polycythemia vera (PV)
Myelodyscrasia (MD)
361
94
496
86
278
Lymphoproliferative
disorders
Malignant lymphomas (ML)
Burkitt's lymphoma (BL)
Acute lymphocytic leukemia (ALL)
Chronic lymphocytic leukemia (CLL)
Monoc1onal gammopathies (MG)
105
24
156
30
23
Solid
tumors
Benign mesenchymal tumors (BMT)
Benign epithelial tumors (BET)
Carcinomas (CA)
Malignant melanomas (MM)
Neurogenic tumors (NT)
Sarcomas (SA)
56
9
87
8
7
7
comes from the published cases in the literature. In addition, there are unpublished cases
from our own laboratory and from other
laboratories, the latter kindly made available
by the original investigators. All in all we have
collected and systematized 1827 cases of
human malignancies exhibiting chromosome
aberrations. This figure does not inc1ude about
2000 cases of chronic myeloid leukemia that
have the Ph1-translocation as the only aberration. According to the diagnoses the material
has been subdivided into 15 c1asses (Table 1).
The aberrations of each patient with a parti-
Table 1. Subdivision of 1827
cases of human neoplasms
with chromosome aberrations
identified by banding techniques
cular disease may be scored and surveyed in
histograms such as the one in Fig. 1. This diagram represents 496 cases of acute myeloid leukemia. We think it must be obvious to all that
the various chromosome types are affected by
aberrations in a nonrandom fashion. Clearly
chromosomes Nos. 7, 8, 17, and 21 are
selectively involved in aberrations.1t should be
noted, however, that the nonrandom elevation
of the aberrations in these chromosomes is
blurred by the random aberrations occasionally affecting all chromosomes and causing the
rather high background level of aberrations.
%
40
30
20+------
10
-1-.....,..--
o
1
3
5
7
9 11 13 15 17 19 21
2
4
6
8 10 12 14 16 18 20 22
Chromosome No.
X
Y
Fig. 1. Histogram of per cent
aberrations affecting the 24 human chromosome types in 496
patients with acute myeloid
leukemia
161
Table 2. Selective involvement of chromosome types in 15 c1asses of human neoplasms
Tumor
type"
Chromosome No.
1 2 3 4
CML
AML
PV
1
5
BMT
BET
CA
MM
NT
SA
1
3
1
1
1
5
3
7
21 22
17
22
8
8
No.of
times
8
involved
a
14
14
14
14
14
7 8
1
1
1
21
20
20 21
9
3
22
17
17
8 9
7 8
8 9
7 8
5
MD
ML
BL
ALL
CLL
MG
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 X Y
14
8
9
22
13 14
-
3
-
3
-
4
8
4
-
1
7
3
2
3
4
Key to abbreviations of the tumor types, see Table 1
D. Clustering of Aberrations to Specific
Chromosomes
Similar histograms have been prepared for all
the different tumor dasses listed in Table 1.
When the chromosomes affected most commonly in each dass are selected, the picture of
Table 2 emerges. It is clear from this table that
not only are aberrations nonrandom within
each tumor class, but there is also a tendency
for the aberrations to cluster to specific chromosomes when different classes are compared.
Thus, chromosomes 1,3,5,7,8,9,14,17,21,
and 22 exhibit nonrandom involvement in
three or more types of malignancy, whereas 12
chromosome types never show any indication
of a selective involvement.
These data from the human tumors are
corroborated and strengthened when similar
data from experimental tumors are taken into
account as weIl. Thus, sarcomas of the rat, the
mouse, and the Chinese hamster all show
involvement of specific chromosomes. The
same is true of rat leukemias and is especially
consistent in mouse leukemias. Virtually
100 % of mouse T -cellleukemias exhibit triso162
my No. 15 (Wien er et al. 1978a,b; Chan et al.
1979).
E. The Significance of Chromosome
aberrations in Malignancy
- A Hypothesis
Thus, banding studies have conclusively demonstrated that chromosome aberrations in
tumors are nonrandom. Unfortunately, this
knowledge does not immediately throw light
on the question what significance this nonrandom involvement has to tumor initiation and
development. Levan and Mitelman (1977)
proposed a model of interpretation, which may
still be useful as a working hypothesis (Fig. 2).
According to this model there are two kinds of
chromosome changes with different modes of
generation: (1) primary or active changes,
which arise through direct interaction between
the oncogenic agent and the genetic material
of the cell, and (2) secondary or passive
changes, which arise at random in the proliferating cell population and are subject to
subsequent selection. The effect of the primary
®
NORMAL
CELL
<:::/"
TRANS FORMED
PREMALIGNANT
CELL
Prlmary event(s): Action of oncogenlc factor(s)
®
Secondary event(s):Numerical and/or structural
chromosome aberrations
CD@@
trlsomy
duplicatlon
translocation
slster
chromosome
exchange
INCREASE IN AFFECTED GENETIC MATERIAL
monosomy
deletion
translocatlon
DECREASE IN UNAFFECTED HOMOLOGOUS
GENETIC MATERIAL
Fig. 2. Hypothetical scheme of chromosomal events in the origin and development of cancer
changes is to transform the cell into an
autonomous parasite and the effect of the
secondary changes is to stepwise ren der the
parasitic cell population more and more malignant. In the model it is assumed that the
primary changes usually are submicroscopic
somatic mutations. Secondary changes cause
chromosome imbalance to amplify the primary
changes and arise from structural rearrangements and nondisjunction.
F. Somatic Mutation in the Initiation of
Malignancy
There are still considerable differences in
opinion about the role of somatic mutation in
cancer. Even though no direct evidence exists
that somatic mutation constitutes the initial
step in oncogenesis, many facts are compatible
with this assumption. Strong support comes
from the fact that most carcinogens have
proven to be potent mutagens.
The work of Knudson in human hereditary
cancers also tends to support the somatic
mutation theory (Knudson et al. 1975). In
retinoblastoma, both sporadic and hereditary
cases exist. According to Knudson's hypothesis two specific somatic mutations in homologous loci are required for a tumor to develop.
This highly unlikely event must be very rare
and is the cause of the low incidence of
sporadic retinoblastoma. In hereditary retinoblastoma, on the other hand, the first mutation
is inherited and thus present in advance in all
somatic cells. Due to the great number of cells
involved in the development of the retina, the
prob ability is now high that the second mutation will take place in at least one cell, making
this cell fully malignant. This hypothetical
scheme is supported by the fact that sporadic
cases virtually always are unifocal, whereas the
majority of the hereditary cases are multifocal.
Recently, some hereditary cases have been
studied where one of the mutations apparently
is substituted for by small chromosomal deletions (Nove et al. 1979). These deletions all
affect chromosome 13 and when different
cases are compared the common segment lost
is band 13q14. It is weIl known from genetic
studies that expression of a recessive mutant
phenotype may be achieved either by a second
163
mutation in the homologous locus or by
adeletion affecting the same locus.
The conclusion from the discussion so far is
that one or more somatic mutation is the
initiating step in at least some tumors. The
somatic mutation may be substituted for by
adeletion and possibly also by other specific
chromosomal aberrations. It is characteristic
of the primary chromosome changes that they
are quite specific and directed towards one or
more genes probably concemed with growth
regulation and differentiation in the affected
tissue.
G. The PhI Chromosome - an Example
of a Primary Chromosome Aberration
The chromosomes of an ample number of
cases of PhI_positive human chronic myeloid
leukemia (CML) have been studied with modern banding techniques. The total number of
cases is not known, since many of them have
not been published, but they must amount to
several thousands. The great majority of them
show one very specific aberration: a translocation between chromosome Nos. 9 and 22.
Usually, patients with this aberration alone are
clinically in the chronic phase of the disease
- when the blastic crisis sets in, secondary
chromosome aberrations ordinarily develop.
Thus CML quite nicely follows the scheme:
a primary specific change followed by secondary less specific changes, a process which is
parallelled by the development of the disease
from a less malignant to a highly malignant
state. It is, of course, possible to hypothesize
that the PhI_positive cells arise not through the
action of some specific mechanism but through
selection from a large number of equally
frequent translocations. There is, however,
overwhelming evidence against this interpretation. A number of PhI_positive leukemias
have been detected in which the translocation
is not the typical t(9 ;22) but so me other
change (Table 3). In all cases one of the
translocation partners is the deleted No. 22,
but the other is not No. 9 but either another
chromosome or the translocation is complex
involving more than two chromosomes. There
are also three cases where Ph 1 appears to
represent a true deletion. All these chromosomal variants are associated with a disease
indistinguishable from typical t(9 ;22) CML.
The first conclusion from these data is that the
significant change in CML is the deletion of
No. 22. The second is that selection from
a random population of translocations is excluded. Obviously, the t(9 ;22) is immensely
overrepresented and other translocations leading to del(22) underrepresented. The conclusion must be that t(9 ;22) in CML is a chromo-
Table 3. Aberrant translocation patterns in 94 cases of PhI-positive CML
No.of
cases
Translocation
No.of
cases
t(X;9;22)
1
t(X;9;17;22)
1
t(1;9 ;22)
4
4
7
t(l ;4;20;22)
1
t(3;4;9 ;22)
t(4;9;17;22)
t(7;9;11;22)
t(9;13 ;15 ;22)
t(9 ;16 ;17 ;22)
t(9;10 ;15 ;19 ;22)
1
1
1
1
1
1
deI (22)
3
No.of
cases
Translocation
t(X;22)
1
t(2 ;22)
2
t(3 ;22)
t(4;22)
t(5 ;22)
t(6;22)
t(7 ;22)
t(9;22)ab
t(10 ;22)
t(ll ;22)
t(12 ;22)
t(13 ;22)
t(14 ;22)
t(15 ;22)
t(16 ;22)
t(17;22)
t(19 ;22)
t(21 ;22)
t(22;22)
1
1
1
2
2
1
2
2
4
1
3
2
3
6
4
1
1
t(2;9 ;22)
t(3;9 ;22)
t(4;9;22)
t(5;9;22)
t(6;9;22)
t(7 ;9;22)
t(8;9 ;22)
t(9;10;22)
t(9;11 ;22)
t(9;13;22)
t(9;14;22)
t(9;15;22)
t(9 ;17 ;22)
t(17 ;17 ;22)
t(21 ;22;22)
Translocation
164
2
2
3
4
1
2
3
2
3
1
2
1
1
some change of the primary type, induced by
a highly specific mechanism which is slightly
error prone and occasionally produces an
aberrant translocation.
H. Secondary Chromosome Aberrations
Are Due to Random Changes and
Subsequent Selection
The cell that has been transformed by a primary change into a premalignant or malignant
state begins to divide without being restricted
by the homeostasis. Secondary chromosome
changes will occur in a random fashion due to
the mechanisms frequent in transformed cells,
i.e. nondisjunction and chromosome breakage
and reunion. Most abnormal cells thus originated will be at selective disadvantage and never
attain any importance in the population. If,
however, a secondary change occurs which
amplifies the effect of the primary chance, the
cell thus changed will be at selective advantage
and divide faster than the surrounding cells.
Since cell division is aprerequisite for secondary changes to become manifest, further
changes will have an increased change of
arising. In this way increased division rate will
generate cells with even more increased division rate in the mode of a vicious circle.
It is this process, we feel, that leads to the
nonrandom distribution of secondary chromosome changes mentioned in the beginning of
the paper. Conversely, it is possible to deduce
from the distribution of aberrations which
chromosomes were originally affected by a primary change. Thus, it is significant that the
most common secondary change in CML is the
duplication by nondisjunction of the Ph1-chromosome.
It has been long recognized that malignant
cell populations give the impression of existing
in a condition of perpetual change and selection. Direct evidence that this may involve
genetic responses leading to the amplification
of specific genes is scarce. Proof that such
a mechanism can be at work has been obtained
in certain experiments of Green et al. (1971).
Growth of human-mouse cell hybrids, selected
in the HAT medium, is dependent on the
human TK + gene situated on the No. 17
chromosome. Normally, such a hybrid will
carry only one No. 17 chromosome. Green and
collaborators drastically decreased the thymi-
dine concentration in the HAT medium and
then selected clones with the best growth
potential. These clones proved to contain both
two and three copies of the TK + chromosome. The obvious conclusion of this experiment
was that this system selected for randomly
occurring nondisjunction products in the cell
population and may thus serve as a model for
the appearance of secondary changes in malignant development.
I. Is it Reasonable to Postulate That Both
Significant and Insignificant
Chromosome Aberrations may be
Found in Neoplastic Cells?
A fact, which has seemed very difficult to
reconcile with the thought of specific and
significant chromosome involvement in the
origin and early development of tumors, is the
presence of a background level of random
aberrations in many materials. It does not
appear biologically reasonable that both significant and insignificant aberrations should be
manifest in tumor cell populations. Recent
work by Wiener and collaborators has thrown
light on this question. In aseries of elegant
experiments these workers have shown that
mouse T -cellleukemias very specifically exhibit No. 15 trisomy (Wiener et a1. 1978a,b).
Furthermore, by using four strains of Robertsonian translocation mice (2n= 38) with
rob(1 ;15), rob(4;15), rob(5 ;15) and rob
(6;15), respectively, they were able to show
that T -cell leukemias in these animals display
trisomy for the entire translocation chromosome (Spira et a1. 1979). Clearly, these neoplasms require trisomy 15 for their development and are not bothered by carrying trisomy
for an insignificant chromosome as well, as
long as the required No. 15 trisomy is present.
We feel that this is a general feature of tumor
cell populations that may explain the high level
of background aberrations in some tumors.
Acknowledgements
Financial support of this work by grants from the
Swedish Cancer Society, the lohn and Augusta
Persson Foundation of Medical Research, and
CANCIRCO is gratefully acknowledged.
165
References
Chan FHP, Ball JK, Sergovich FR (1979) Trisomy
15 in murine thymomas induced by chemical carcinogens, X-irradiation, and an endogenous murine
leukemia virus. J Natl Cancer Inst 62:605-610
- Green H, Wang R, Kehinde 0, Meuth M (1971)
Multiple human TK chromosomes in human-mouse
somatic cell hybrids. Nature (London) New Biol
234: 138-140 - Knudson AG, Heathcote HW,
Brown BW (1975) Mutation and childhood cancer:
A probabilistic model for the incidence of retinoblastoma. Proc Natl Acad Sci 72:5116-5120 - Levan
G, Mitelman F (1977) Chromosomes and the
etiology of cancer. In: Chapelle A de la, Sorsa
M (eds) Chromosomes today, vol6. Elsevier/NorthHolland, Amsterdam, pp 363-371 - Nove J, Little
166
JB, Weichselbaum RR, Nichols WW, Hoffman
E (1979) Retinoblastoma, chromosome 13, and in
vitro cellular radiosensitivity. Cytogenet Cell Genet
24: 176-184 - Spira J, Wiener F, Ohno S, Klein
G (1979) Is trisomy cause or consequence ofmurine
T -cellieukemia development? Studies on Robertsoni an translocation mice. Proc Natl Acad Sci
76:6619-6621 - Wiener F, Ohno S, Spira J,
Haran-Ghera N, Klein G (1978a) Chromosome
changes (trisomies *15 and 17) associated with
tumor progression in leukemias induced by radiation
leukemia virus. J Natl Cancer Inst 61 :227-237
- Wiener F, Spira J, Ohno S, Haran-Ghera N, Klein
G (1978b) Chromosome changes (trisomy 15) in
murine T-cellleukemia induced by 7,12-dimethylbenz(a)anthracene (DMBA). Int J Cancer
22:447-453